Will the cashier be the first occupation to be replaced by artificial intelligence?

With the continuous development of new technology research and technical barriers, AI's vents have arrived. Commercial applications of AI are spreading across all areas, and retail is no exception. YI Tunnel, a company that applies AI technology to the traditional retail industry, has launched an AI smart cashier settlement robot and expects it to replace the simple repetitive work of the cashier in the future.

Project Highlights:

YI Tunnel is a combination of artificial intelligence hardware and retail startups, in short, replacing the recruiters with artificial intelligence cashier robots. YI Tunnel uses Convolutional Neural Networks to import retail merchandise libraries and optimize the robots to optimize the artificial cash register.

YI Tunnel has completed the testing phase and will be put into use.

Japan's small convenience stores and Amazon Go have made corresponding attempts to smart cashiers. The former uses RFID-embedded RFID technology, which uses visual recognition technology and deep algorithm learning technology to achieve unattended cash. Not put into mass use.

The data shows that China will also face the problem of population aging and shortage of labor in the future. Like Japan, simple repetitive work will be replaced by machines more and more, and YI Tunnel looks forward to exploring the business opportunities that this trend will bring.

"Today's AI is moving far faster than our fantasies." On May 22, Ke Jie wrote this sentiment the night before the match with AlphaGo.

This is indeed the case. With the continuous development of new technologies and technical barriers, the enthusiasm of human intelligence (AI) has arrived. "The combination of smart hardware vertical and subdivided verticals will be one of the future trends," said Jiang Yujie, a partner at Xintian Ventures, who also publicly expressed his optimism and focus on the combination of smart hardware and verticals. The commercial application of artificial intelligence is spreading quietly in various fields, accelerating the extension of artificial intelligence to various vertical fields, and the retail industry is no exception.

YI Tunnel is a company that applies artificial intelligence technology to the traditional retail industry and commercializes artificial intelligence and retail scenarios. YI Tunnel has launched an AI smart cashier settlement robot and expects this robot to replace the simple repetitive work of the cashier in the future.

Business Opportunity: Based on the refinement of smart sensor technology

"With the improvement of sensor input quality and input precision of intelligent hardware, sensor prices are declining, technical barriers are constantly being broken, and intelligent robots can achieve more and more things." Chen Qi Capital Investment Director Chen Wei in the interview “With the coming labor shortage and aging population in China, simple and repetitive labor will be replaced.”

In terms of technology, the artificial intelligence visual recognition technology is continuously improved and iterated, and the intelligent robot image recognition sensor is continuously refined, which provides a technical basis for the YI Tunnel cash register robot. In terms of visual technology, foreign countries currently mainly adopt the residual network architecture. In the 2015 ImageNet Challenge, the 152-layer network completed the 3% error rate challenge, which reduced the error rate of machine-to-entity recognition to a very low level. This matter provided great confidence for Wu Yili's early entrepreneurial ideas.

On this basis, the YI Tunnel cashier robot splits the object classification into several steps, which are completed by using different residual networks from tens of layers to hundreds of layers, including object pull frame, object recognition, and camera voting. The error rate was further reduced from 3% in the academic world to 0.3%, which is about 1/5 of the artificial error rate.

In the business context, the replacement of artificially intelligent robots for simple repetition or even night shift work is one of the future trends. The YI Tunnel intelligent robot combines artificial intelligence with the closing scenario of the retail industry to develop a smart settlement robot.

YI Tunnel saw a pain point in the retail industry. “Either large supermarkets, chain convenience stores and small community stores require a lot of manpower to check out the cashier. 24-hour convenience stores need to pay more expensive night-time cashier fees, peak merchants The ultra-queuing phenomenon is also serious," said Liu Juan, marketing director of YI Tunnel. "YI Tunnel can complete the simple repetitive work of checkout, release the labor force and enhance the consumer experience."

Core Competence: Visual Recognition Technology of Convolutional Neural Networks

YI Tunnel uses convolutional neural network technology to teach computers to identify and judge product images.

The convolutional neural network technology is somewhat like the process of teaching children to recognize objects. The computer is programmed to teach the computer to learn graphics and shapes. As the picture base of the input computer continues to increase, the accuracy of the computer's image recognition will also increase.

"In the past, if computers wanted to know Apple, they needed to summarize all the characteristics of Apple to tell the computer accurately, such as 6 to 8 cm in diameter, red, oval, etc., but the real apples are different, so the error rate is very high. Gao." Wu Yili, founder of YI Tunnel, said, "When the convolutional neural network is identified, we will look at the pictures of thousands of apples for the computer. The computer will learn about Apple itself, and then whatever Apple has. Changes, computers can be recognized as accurately as people, this method can reduce the error rate."

So in addition to the algorithm, the amount of data is also critical for YI Tunnel. At present, YI Tunnel's material library has entered and identified 3,000 SKUs, 2 million photos, and expanded to 30,000 SKUs in the next year, marking more than 30 million photos for more accurate product identification.

YI Tunnel can identify vegetables, fruits, retail, beverages, alcohol, stationery, clothing, toys and other commodities. The technology can be applied to retail closing scenarios such as large supermarkets, convenience stores, pharmacies and bookstores.

When the reporter tried YI Tunnel, YI Tunnel not only recognized the difference between Apple and other fruits, but also automatically weighed, calculated the price, and recognized apples and snakes. However, if you try something that is not recorded, YI Tunnel is not identifiable, such as iPhone, packet napkins and wallets.

According to the data provided by the YI Tunnel, the time for the YI Tunnel intelligent settlement machine to identify the product is about 0.02 seconds, which is faster than the manual scan code recognition. Moreover, the error rate of traditional manual checkout is about 1.5%. In the test phase, YI Tunnel can reduce the error rate to less than 0.3%, which is one-fifth of the traditional manual settlement error rate.

When Shangchao needs to update new products, YI Tunnel can upload the graphics and pictures of the products to the visual recognition effect, and introduce the new products into the settlement robot of Supermarket. “You can understand that just like the APP Store update application, after we write new products at the headquarters, the YI Tunnel settlement machine held by the merchant only needs to update the system to settle the newly added goods.” Marketing Director of YI Tunnel Liu Juan said.

At present, there are more than 40 people in the YI Tunnel team, including about 6 core technical teams, mainly from the Tsinghua alumni of founder Wu Yili. Wu Yili also graduated from Tsinghua University and worked for Oracle and IBM's e-commerce division.

There are about 30 engineers in the YI Tunnel team working in the Sichuan branch, mainly responsible for data collection and data marking. These engineers completed the programming of 2 million images in 8 months, which is to teach intelligent robots to recognize and recognize the images of retail products.

Market Competitor: Amazon Go and other unoccupied cashier companies

In fact, replacing the simple and repetitive labor force YI Tunnel in the retail industry with intelligent robots or other artificial intelligence methods is not a precedent. Other companies have tried before the YI Tunnel.

At the end of 2016, Amazon officially released a 1 minute 49 second promotional video on Youtube, confirming the launch of the new concept retail store Amazon Go. Amazon Go is an open, unmanned retail store that abandons the traditional supermarket's cashier checkout process, so customers no longer have to carry a shopping basket to check out, just pick up what you want and then walk out of the store.

Amazon Go integrates a large number of sensors that recognize people's movements, merchandise, and merchandise locations to complete the unfinished process. Amazon Go uses visual recognition technology, deep learning algorithms, and sensor fusion technology to achieve unmanned shopping. Amazon Go is expected to open in Seattle early this year, but it is not yet open and is now only open to Amazon employees.

Five large convenience stores such as Japan 7-11 have also proposed to introduce an unmanned cash register system before 2025, and introduce self-checkout counters for consumers to check out their own in all stores in Japan. Japan's unmanned cash register system is based on IC tags embedded in commodity packaging. IC tags are less than 1 mm thick. Using technology called RFID (radio frequency identification), product information can be written and information can be read by machine. .

According to the founding team of YI Tunnel, most unmanned convenience stores use RFID technology, and the technology still has certain barriers, for example, it still requires high labor loss. The YI Tunnel uses a convolutional neural network to teach computers the ability to learn about retail goods, compared to Amazon's technology model.

In addition, there are already some early unoccupied cash registers in China, which are generally in the stage of development, and most of them also use RFID technology. At present, the relatively closed mode is adopted, and only one person can enter only one person at a time.

The future: a business model that is profitable with hardware

YI Tunnel has passed the testing phase and is ready to enter the market for use.

YI Tunnel hopes to start using products in a small scale and cooperate with some mutually recognized brands, while continuing to iteratively optimize YI Tunnel's intelligent robot products. Therefore, YI Tunnel will adopt a model of cooperation with small-scale concept stores and flagship stores this year to jointly shape the brand.

YI Tunnel is able to collect product data, sales data, and other store-related data while reducing the operating costs of retail stores by replacing the cashier's simple repetitive work. In addition to completing hardware services, some data support services are not excluded in the future. For example, through YI Tunnel's cashier system, chain store owners can see the sales of all storefronts and a single storefront. This data can be updated in real time to facilitate retail shipments and replenishment.

In the future, YI Tunnel will realize profit by selling a combination of intelligent cash register hardware and data processing. YI Tunnel is cooperating with two domestic supermarket chains and a chain convenience store. In addition, YI Tunnel can also pass the check-and-pay mode, so that the machine can enter the mall for free to achieve a win-win situation.

Up to now, YI Tunnel has not completed any financing, and the main R&D expenses come from the seed round funded by the founding team itself. “The reason why it is named YI Tunnel is that we want this smart hardware to be called a tunnel for link technology and business.” Wu Yili, the founder of the company, believes that “although Amazon and Japanese convenience stores already have similar products, YI Tunnel is the first to be put into business. Practical products."

Artificial intelligence and intelligent hardware are just beginning to develop in the vertical field. Whether the commercialization of intelligent hardware in the vertical field can be achieved depends on the technical capabilities of the founding team and the commercial perception of the founders. This is also Li Kaifu’s repeated emphasis on finding China’s makers conference and GMIC conference: “There are often some excellent artificial intelligence scientists who only research, business sense is not good enough, and people who understand business are not necessarily good, if the two are combined, then It will be perfect."

There is still a long way to go before the commercialization of artificial intelligence, but as the progress of individual projects and technical barriers have been broken, the field has taken the first step.


New Toys

Familie doll has a blank face,you can draw eyes,lips,nose,eyebrows on the face and you can sew according to your drawing.You can also send us the picture of your drawing and we can sew for you.It likes a mystery box,you don't know what it is until you recieve Familie doll.

New Toys,New Toy Doll,New Barbies Toy,New Doll Set

Nantong Einsame Familie Toy Technology Co.,Ltd , https://www.einsamefamilies.com